The post The Big Question (Statistical Analysis of Climate Extremes: The Blog about the Book. Part 4) first appeared on FifteenEightyFour | Cambridge University Press.

]]>My preferred approach is based on the ideas of the philosopher Plato. That means, assume the existence of a climate reality in space and time, a reality also about the big climate question. The truth is not exactly known to us, but we can approach it by means of data and statistical analysis. We make an inference of climate reality with uncertainties. We report our results with error bars. We write as clearly as possible (see Part 3 of the blog). We support readers with the data and the software used for the analyses. This helps them to replicate and understand what we did. The climate research community proceeds jointly to get closer to the climate reality. Indeed I have the optimism that humans can learn.

More names on the scientific method? Take Immanuel Kant or Karl Popper. Ignore the philosophers who think it is fancy to speculate whether truth exists. Or whether it is a mere social construct. Physics, the climate system and big question on climate extremes are rather unimpressed by such attempts. A couple of years ago I posted a blog series (in German) on floods and climate at some German climate blog. I got a few critical remarks by sociologists. No problem, just ignore such.

Fine. In the book, after the reader has been introduced to the big question (Chapter 1), the first to present were the data (Chapter 2) used to illustrate the statistical methods (Chapter 3). Let us ignore the formulas here. The major target of the analyses is the rate of occurrence of extremes per time unit. For example, how many hurricanes of a certain category happened per season?

Since climate changes with time, we have to admit a time-dependent occurrence rate. The uncertainties in occurrence-rate estimation are obtained by computing-intensive mathematical simulations. That means, use a random number generator and produce artificial series of events, which you then re-analyse for the occurrence rate. You repeat this exercise many times. From the many simulated occurrence-rate curves, you can then construct confidence bands. For example, a 90% confidence band comes from discarding the upper 5% of the simulation values and the lower 5%. This is called a percentile confidence band. (For the expert: the book advocates the usage of a more sophisticated, high-accuracy band of type percentile-t.) A typical result curve looks like this.

What do we see here? The horizontal axis is time, the past four decades. The vertical axis is occurrence rate of major hurricanes (categories 4 to 5) making landfall in the United States of America. The units are one over years. Conventionally, we use the Greek symbol λ (lambda). The “hat” on top of λ indicates that this is an estimate. The solid line is the estimated time-dependent occurrence rate curve. The shaded area is a 90% percentile-t confidence band. The true occurrence rate at a certain time is with a chance of 90% within the shaded band. The major result is an increasing trend in hurricane occurrence. The current value is about 0.2 one-over-years units, that means, one major event every fifth year. This means on average such an event happens, not exactly every fifth year (then hurricane risk prediction would be fairly easy). Although the uncertainties are pretty large, the trend is there.

My analysis job is done at this stage. The methods the textbook presents allow the calculation and the construction of such time-dependent risk curves. Analysts may use the risk curves as a basis for designing reinsurance products. Meteorologists may use the curves to test hypotheses about driving factors. Climate researchers have another curve demonstrating the impacts of climate change. Stakeholders may make better-informed decisions. To repeat, the research community proceeds jointly to get closer to the reality about the time-dependent risk of hurricanes and other types of extremes. These will be covered in the next blog article.

The book published by Cambridge University Press in May 2020. It is available to order in print, online for individuals or online for institutions.

Online material for the book: https://www.manfredmudelsee.com/textbook/index.htm

Hochwasserrisiko: Persönliche Anmerkungen von Manfred Mudelsee, Blog „Die Klimazwiebel“ (In German) https://klimazwiebel.blogspot.com/2011/12/hochwasserrisiko-personliche.html (last access 15 June 2020)

Plato picture, excerpt from a painting from circa 1325 to 1335

https://www.gla.ac.uk/myglasgow/library/files/special/images/chaucer/H231_0276wf.jpg (last access 15 June 2020)

Statistical Analysis of Climate Extremes: The Blog about the Book. Part 1: Corona

https://www.cambridgeblog.org/2020/05/statistical-analysis-of-climate-extremes-the-blog-about-the-book-part-1-corona/

Statistical Analysis of Climate Extremes: The Blog about the Book. Part 2: The Cover

https://www.cambridgeblog.org/2020/05/statistical-analysis-of-climate-extremes-the-blog-about-the-book-part-2-the-cover/

Statistical Analysis of Climate Extremes: The Blog about the Book. Part 3: Content

https://www.cambridgeblog.org/2020/06/statistical-analysis-of-climate-extremes-the-blog-about-the-book-part-3-content

The post The Big Question (Statistical Analysis of Climate Extremes: The Blog about the Book. Part 4) first appeared on FifteenEightyFour | Cambridge University Press.

]]>The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 3: Content first appeared on FifteenEightyFour | Cambridge University Press.

]]>The audience comprises three more types of people. First, the university teachers of the students have to understand and get into the subject. Perhaps they will start teaching a course on climate extremes this year? Note that there is a PDF with the solutions to the exercises available to registered teachers. Second, risk analysts in the insurance industry may have to quickly learn methods and the state of current knowledge about climate extremes. Those guys I brief with a “Summary for Risk Analysts” in each of the application chapters (Floods and Droughts, Heatwaves and Cold Spells, Hurricanes and Other Storms). Third, professional researchers also benefit from the book because of the supplied software tools and the “Outlook” sections, where the current state of knowledge and future directions are explored.

The structure follows well-proven paths: Introduction, Data, Methods and then the three application chapters. I gave many appendices to make the book as self-contained as possible and meaningful: measurements, climate archives, climate models, statistical inference, numerical techniques, data, software and a glossary. Honestly, I believe that also undergraduate students with only a basic training in calculus and statistics can do it. Most important is an open mind and the joy about learning something new.

The full contents list can be seen here.

With respect to the act of writing, I started with Introduction, Data and Methods. Then the appendices. This allowed to keep the application chapters – fields where new research is published daily – fairly recent. (I sent the book files in summer 2019.)

As regards the breadth, I realized that I should not compete with the IPCC, which delivers in its Assessment Reports a more or less full coverage of the weather–climate research field (also at a fair recentness). This led me to present the application of the methods to selected examples of climate extremes in case studies. Step-by-step explanation of the analysis. The interplay between data, questions to the data, the search for suitable tools for obtaining an answer and the interpretation of the results. I believe that this style serves best to achieve the goal of the book.

The book, published by Cambridge University Press in May 2020. It is available to order in print, online for individuals or online for institutions.

Online material for the book: https://www.manfredmudelsee.com/textbook/index.htm

The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 3: Content first appeared on FifteenEightyFour | Cambridge University Press.

]]>The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 2: The Cover first appeared on FifteenEightyFour | Cambridge University Press.

]]>Photo selection is an intimate matter. Your soul kind of forms a synapse that is open to other souls. Ethical values learned as a child and your aesthetic education enter the arena. Your ratio is rather unimportant, but not turned off.

On the search for a cover picture, my ratio told me to look for red. One reason was pragmatic: on the shelf, the new book should look differently than my previous book on climate and time series, which has a blue cover. Red, for me, also transports a feeling for the danger our climate system is in.

But our mother nature, the undisturbed system, holds a rich beauty. This has to be shown as well. The next ingredient of my filter was preciseness. The picture must not blur the danger nor the beauty. It must correspond to the pursued clarity of the presentation of the mathematical risk formulas in the book. Everyone should understand.

When I first saw the initial selection of my search for “red” and “nature” on Getty Images, I immediately felt that Dhwee’s lagoon photo was it. As a matter of decency, I added a few other pictures and showed the final selection to my wife. It was clear to me that she agreed on the lagoon.

Cambridge then did a great job with the finishing of the cover. No new colours, just red, black and white. The sharpness of the Futura typeface (especially sensible for the “M” of which my name possesses two exemplars) – it reminds me of Wally Broecker’s warning that we humans should not provoke the angry climate beast with greenhouse gases. Fantastic!

Once finished, I am already exploring in my mind the next book project. Blue, red – green! Something with ecology and natural life, somehow analysed by means of mathematical techniques? Give me some time. Let me in this blog series talk about the red book.

Broecker WS (1995) The glacial world according to Wally. Eldigio Press, Palisades, New York, 318 pp. [the 2002 version is available at https://www.ldeo.columbia.edu/~broecker/Home_files/GlacialWorld.pdf]

Mudelsee M (2014) Climate Time Series Analysis: Classical Statistical and Bootstrap Methods. Second edition. Springer, Cham, Switzerland, 454 pp.

Cambridge University Press is publishing a book of mine on climate extremes. It comes out in May 2020. It is available at to pre-order in print, online for individuals or online for institutions.

Online material for the book: https://www.manfredmudelsee.com/textbook/index.htm

The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 2: The Cover first appeared on FifteenEightyFour | Cambridge University Press.

]]>The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 1: Corona first appeared on FifteenEightyFour | Cambridge University Press.

]]>The common thing between Corona and climate extremes is the risk put on our society and economy. Risk is defined as the probability that within a time interval something bad happens, for example, someone dies from or a hurricane occurs.

The data for my book consist in the dates when the bad events happened, such as 29 August 2005, when hurricane Katrina peaked. Sometimes there is more data information, such as the maximum wind speed. Besides wind speed, the book is also about temperature extremes (heatwaves or cold spells) and precipitation extremes (floods and droughts). These three are the most relevant climate variables. The big question is how global climate change influences the occurrence of climate extremes. This plagues decision-makers as well as climate researchers.

According to Cambridge’s website, my book gives “an accessible overview of the statistical analysis methods which can be used to investigate climate extremes and analyse potential risk. The statistical analysis methods are illustrated with case studies. The book also provides datasets and access to appropriate analysis software, allowing the reader to replicate the case study calculations.” I may only add that you can use the software also to analyse your own data. There is no technical hindrance for becoming a climate risk analyst. Or a health risk analyst, if you apply my methods to medical data.

Hopefully other researchers find the book and the software tools useful. It would be great if the tools would enable them to obtain results that are of relevance to society and economy. This little blog series is about various facets of the book. This should help you to better know about the product, whether you are inclined to buy it. I would be glad to respond to your queries on the book.

Big changes in the world also offer to learn rapidly, quasi by force. Corona taught me how important it is for my company to offer the courses (on climate time series and risk analysis) also online. So, at the moment I am learning the hardware and software aspects of making video tutorials.

I am an optimistic person. I saw the fast and drastic reactions of the decision-makers on the Corona risk, in my country and worldwide. I sincerely believe that mankind has the ability to react adequately also to the risks imposed by climate change. In this area, we have more time than now on the virus: slightly more. This must not make us say “after the virus”. The next one will come. We have to live with both. Viruses and climate change.

Cambridge University Press is publishing a book of mine on climate extremes. It comes out in May 2020. It is available at to pre-order in print, online for individuals or online for institutions. Online material for the book is at my own academic website, https://www.manfredmudelsee.com/textbook/index.htm.

The post Statistical Analysis of Climate Extremes: The Blog about the Book. Part 1: Corona first appeared on FifteenEightyFour | Cambridge University Press.

]]>